Thesis (M.A.)--Georgia State University, 2005. / Title from title screen. Jeremy Crampton, committee chair; Michael Eriksen, Dona Stewart, committee members. Electronic text (94 p. : ill., maps (some col.)) : digital, PDF file. Description based on contents viewed July 10, 2007. Includes bibliographical references (p. 88-94).

Disease mapping and spatial statistics have become an important part of modern day statistics and have increased in popularity as the methods and techniques have evolved. The application of disease mapping is not only confined to the analysis of diseases as other applications of disease mapping can be found in Econometric and financial disciplines. This thesis will consider two data sets. These are the Georgia oral cancer 2004 data set and the South African acute pericarditis 2014 data set. The Georgia data set will be used to assess the hyperprior sensitivity of the precision for the uncorrelated heterogeneity and correlated heterogeneity components in a convolution model. The correlated heterogeneity will be modelled by a conditional autoregressive prior distribution and the uncorrelated heterogeneity will be modelled with a zero mean Gaussian prior distribution. The sensitivity analysis will be performed using three models with conjugate, Jeffreys' and a fixed parameter prior for the hyperprior distribution of the precision for the uncorrelated heterogeneity component. A simulation study will be done to compare four prior distributions which will be the conjugate, Jeffreys', probability matching and divergence priors. The three models will be fitted in WinBUGS® using a Bayesian approach. The results of the three models will be in the form of disease maps, figures and tables. The results show that the hyperprior of the precision for the uncorrelated heterogeneity and correlated heterogeneity components are sensitive to changes and will result in different results depending on the specification of the hyperprior distribution of the precision for the two components in the model. The South African data set will be used to examine whether there is a difference between the proper conditional autoregressive prior and intrinsic conditional autoregressive prior for the correlated heterogeneity component in a convolution model. Two models will be fitted in WinBUGS® for this comparison. Both the hyperpriors of the precision for the uncorrelated heterogeneity and correlated heterogeneity components will be modelled using a Jeffreys' prior distribution. The results show that there is no significant difference between the results of the model with a proper conditional autoregressive prior and intrinsic conditional autoregressive prior for the South African data, although there are a few disadvantages of using a proper conditional autoregressive prior for the correlated heterogeneity which will be stated in the conclusion.

Understanding spatial perspectives on the spread and incidence of a disease is invaluable for public health planning and intervention. Choropleth maps are commonly used to provide an abstraction of disease risk across geographic space. These maps are derived from aggregated population counts that are known to be affected by the small numbers problem. Kernel density estimation methods account for this problem by producing risk estimates that are based on aggregations of approximately equal population sizes. However, the process of aggregation often combines data from areas with non-uniform spatial and population characteristics. This thesis presents a new method to aggregate space in ways that are sensitive to their underlying risk factors. Such maps will enable better public health practice and intervention by enhancing our ability to understand the spatial processes that result in disparate health outcomes.

Disease maps are powerful tools for depicting spatial variations in disease risk and its underlying drivers.  However, producing effective disease maps requires careful consideration of the statistical and spatial properties of the disease data. In fact, the choice of mapping method influences the resulting spatial pattern of the disease, as well as the understanding of its underlying population characteristics. New developments in mapping methods and software in addition to continuing improvements in data quality and quantity are requiring map-makers to make a multitude of decisions before a map of disease burdens can be created. The impact of such decisions on a map, including the choice of appropriate mapping method, not been addressed adequately in the literature. This research demonstrates how choice of mapping method and associated parameters influence the spatial pattern of disease. We use four different disease-mapping methods – unsmoothed choropleth maps, smoothed choropleth maps produced using the headbanging method, smoothed kernel density maps, and smoothed choropleth maps produced using spatial empirical Bayes methods and 5-years of zip code level HIV incidence (2007- 2011) data from Dallas and Tarrant Counties, Texas. For each map, the leading population characteristics and their relative importance with regards to HIV incidence is identified using a regression analysis of a CDC recommended list of socioeconomic determinants of HIV. Our results show that the choice of mapping method leads to different conclusions regarding the associations between HIV disease burden and the underlying demographic and socioeconomic characteristics. Thus, the choice of mapping method influences the patterns of disease we see or fail to see. Accurate depiction of areas of high disease burden is important for developing and targeting appropriate public health interventions.

Global spread of infectious disease threatens the well-being of human, domestic, and wildlife health. A proper understanding of global distribution of these diseases is an important part of disease management and policy making. However, data are subject to complexities by heterogeneity across host classes and space-time epidemic processes [Waller et al., 1997, Hosseini et al., 2006]. The use of frequentist methods in Biostatistics and epidemiology are common and are therefore extensively utilized in answering varied research questions. In this thesis, we proposed the Hierarchical Bayesian approach to study the spatial and the spatio-temporal pattern of tuberculosis in Kenya [Knorr-Held et al., 1998, Knorr-Held, 1999, L opez-Qu lez and Munoz, 2009, Waller et al., 1997, Julian Besag, 1991]. Space and time interaction of risk (ψ[ij]) is an important factor considered in this thesis. The Markov Chain Monte Carlo (MCMC) method via WinBUGS and R packages were used for simulations [Ntzoufras, 2011, Congdon, 2010, David et al., 1995, Gimenez et al., 2009, Brian, 2003], and the Deviance Information Criterion (DIC), proposed by [Spiegelhalter et al., 2002], used for models comparison and selection. Variation in TB risk is observed among Kenya counties and clustering among counties with high TB relative risk (RR). HIV prevalence is identified as the dominant determinant of TB. We found clustering and heterogeneity of risk among high rate counties and the overall TB risk is slightly decreasing from 2002-2009. Interaction of TB relative risk in space and time is found to be increasing among rural counties that share boundaries with urban counties with high TB risk. This is as a result of the ability of models to borrow strength from neighbouring counties, such that near by counties have similar risk. Although the approaches are less than ideal, we hope that our formulations provide a useful stepping stone in the development of spatial and spatio-temporal methodology for the statistical analysis of risk from TB in Kenya. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

West Nile virus (WNV) is an emerging infectious disease that has widespread implications for public health practitioners across the world. Within a few years of its arrival in the United States the virus had spread across the North American continent. This research focuses on the development of a spatially explicit GIS-based predictive epidemiological model based on suitable environmental factors. We examined eleven commonly mapped environmental factors using both ordinary least squares regression (OLS) and geographically weighted regression (GWR). The GWR model was utilized to ascertain the impact of environmental factors on WNV risk patterns without the confounding effects of spatial non-stationarity that exist between place and health. It identifies the important underlying environmental factors related to suitable mosquito habitat conditions to make meaningful and spatially explicit predictions. Our model represents a multi-criteria decision analysis approach to create disease risk maps under data sparse situations. The best fitting model with an adjusted R2 of 0.71 revealed a strong association between WNV infection risk and a subset of environmental risk factors including road density, stream density, and land surface temperature. This research also postulates that understanding the underlying place characteristics and population composition for the occurrence of WNV infection is important for mitigating future outbreaks. While many spatial and aspatial models have attempted to predict the risk of WNV transmission, efforts to link these factors within a GIS framework are limited. One of the major challenges for such integration is the high dimensionality and large volumes typically associated with such models and data. This research uses a spatially explicit, multivariate geovisualization framework to integrate an environmental model of mosquito habitat with human risk factors derived from socio-economic and demographic variables. Our results show that such an integrated approach facilitates the exploratory analysis of complex data and supports reasoning about the underlying spatial processes that result in differential risks for WNV. This research provides different tools and techniques for predicting the WNV epidemic and provides more insights into targeting specific areas for controlling WNV outbreaks.

Title from title page of PDF (University of Missouri--St. Louis, viewed March 9, 2010). Includes bibliographical references.

In this paper, I will show how applying a spatially explicit context to an existing environmental health surveillance framework is vital for more complete surveillance of disease, and for disease prevention and intervention strategies. As a case study to test the viability of a spatial approach to this existing framework, the risk of human exposure to Lyme disease will be estimated. This spatially explicit framework divides the surveillance process into three components: hazard surveillance, exposure surveillance, and outcome surveillance. The components will be used both collectively and individually, to assess exposure risk to infected ticks. By utilizing all surveillance components, I will identify different areas of risk which would not have been identified otherwise. Hazard surveillance uses maximum entropy modeling and geographically weighted regression analysis to create spatial models that predict the geographic distribution of ticks in Texas. Exposure surveillance uses GIS methods to estimate the risk of human exposures to infected ticks, resulting in a map that predicts the likelihood of human-tick interactions across Texas, using LandScan 2008TM population data. Lastly, outcome surveillance uses kernel density estimation-based methods to describe and analyze the spatial patterns of tick-borne diseases, which results in a continuous map that reflects disease rates based on population location. Data for this study was obtained from the Texas Department of Health Services and the University of North Texas Health Science Center. The data includes disease data on Lyme disease from 2004-2008, and the tick distribution estimates are based on field collections across Texas from 2004-2008.

This thesis addresses three interrelated challenges of disease mapping and contributes a new approach for improving visualization of disease burdens to enhance disease surveillance systems. First, it determines an appropriate threshold choice (smoothing parameter) for the adaptive kernel density estimation (KDE) in disease mapping. The results show that the appropriate threshold value depends on the characteristics of data, and bandwidth selector algorithms can be used to guide such decisions about mapping parameters. Similar approaches are recommended for map-makers who are faced with decisions about choosing threshold values for their own data. This can facilitate threshold selection. Second, the study evaluates the relative performance of the adaptive KDE and spatial empirical Bayes for disease mapping. The results reveal that while the estimated rates at the state level computed from both methods are identical, those at the zip code level are slightly different. These findings indicate that using either the adaptive KDE or spatial empirical Bayes method to map disease in urban areas may provide identical rate estimates, but caution is necessary when mapping diseases in non-urban (sparsely populated) areas. This study contributes insights on the relative performance in terms of accuracy of visual representation and associated limitations. Lastly, the study contributes a new approach for delimiting spatial units of disease risk using straightforward statistical and spatial methods and social determinants of health. The results show that the neighborhood risk map not only helps in geographically targeting where but also in tailoring interventions in those areas to those high risk populations. Moreover, when health data is limited, the neighborhood risk map alone is adequate for identifying where and which populations are at risk. These findings will benefit public health tasks of planning and targeting appropriate intervention even in areas with limited and poor-quality health data. This study not only fills the identified gaps of knowledge in disease mapping but also has a wide range of broader impacts. The findings of this study improve and enhance the use of the adaptive KDE method in health research, provide better awareness and understanding of disease mapping methods, and offer an alternative method to identify populations at risk in areas with limited health data. Overall, these findings will benefit public health practitioners and health researchers as well as enhance disease surveillance systems.